I mean, thats the way the capitalist, stock-return-driven economy works. The market expects a company to constantly grow to pump their stock price, so they have to find new revenue or cut costs somewhere. But they can't do that forever...
The founders build a great product to pull in users, then they go public, then the MBAs turn to enshittification to drive more revenue and get rich while they can. The rest of us then move on to the next platform, if it even exists....
Yes, and they'll use legislation to pull up the ladder behind them. It's a form of Regulatory Capture, and it will absolutely lock out small players.
But there are open source AI training datasets, but the question is whether LLMs can be trained as accurately with them.
This is awesome, but now we need better battery tech that doesn't rely on lithium and cobalt. Getting that up to this scale will be hard, though.
Ok, that's true
Capitalism promises nothing, my friend.
Isn't this standard enshitification?
Google started off like any new tech company years ago and they did have a superior product. Then they went public and Wall Street started expecting certain revenue numbers every quarter, growth slowed because they saturated the market, so the core product was pushed toward making money instead staying true to the original goals.
Typical stuff, really
So what? The biggest part of picking a President is the people they're going to put in their cabinet and surround themselves with. That's the problem with Trump. He's going to fill the presidency with a bunch of maga ass hats, on top of doing his own damage.
Article says the glass plates can only be written once, so don't toss out your hdd just yet
I don't disagree, but they still need devs to build and maintain the custom functions required. So then they sign up with Red Hat and still pay huge dollars. Unless they hire in house devs, which they rarely seem to want to do (at the levels needed for these projects).
In my field, the way I've seen Oracle contracts rationalized is the same way IBM contracts are rationalized.... they're stupid expensive, usually under deliver, but they're the biggest names. When the project goes south, the buyer can tell their superiors "well we hired IBM, and they're the best, so what else could I have done?" It's a form of of CYA.
I looked it up and you're correct. I didn't realize Uber started literally as "UberCab" and later dropped the "cab" and added the personal car ride sharing component. Thx for the tip!
Thanks for the link to Common Crawl; I didn't know about that project but it looks interesting.
That's also an interesting point about heavily curated data sets. Would something like that be able to overcome some of the bias in current models? For example, if you were training a facial recognition model, access a curated, open source dataset that has representative samples of all races and genders to try and reduce the racial bias. Anyone training a facial recognition model for any purpose could have a training set that can be peer reviewed for accuracy.